Optimum model for predicting temperature settings on hot dip galvanising line F. J. Martı ´nez-de-Piso ´n* 1 , A. V. Pernı ´a 1 , A. Gonza ´lez 1 , L. M. Lo ´ pez-Ochoa 2 and J. B. Ordieres 3 Controlling the annealing cycle in a hot dip galvanising line (HDGL) is vital if each coil treated is to be properly galvanised and the steel is to have the right properties. Current HDGL furnace control models usually take into account the dimensions of the coil to be dipped and, in some cases, the type of steel. This paper presents a new model for monitoring furnace temperature settings, which considers not just the coil dimensions but also the chemical composition of the steel. This enables the model to be adjusted more suitably to each type of steel to be dipped, so that the HDGL annealing cycle is optimised and rendered more efficient in dealing with new products. The ultimate aim is to find a model that is equally efficient for new types of steel coil that have not been processed before and whose dimensions and chemical compositions are different from coils processed previously. To find the best model, this paper compares various new and classical algorithms for developing a precise and efficient prediction model capable of determining the three temperature settings for heating on an HDGL located in Avile ´ s (Spain) on the basis of the physical and chemical characteristics of the coils to be processed and the preset process conditions. Keywords: Hot dip galvanising line, Annealing furnace, Data mining, Artificial intelligence, Modelling industrial processes List of symbols Al, Cu, Ni, Cr, Nb chemical composition of steel, wt-% C, Mn, Si, S, P chemical composition of steel, wt-% THC1 zone 1 set point temperature (initial heating zone), uC THC3 zone 3 set point temperature (intermediate heating zone), uC THC5 zone 5 set point temperature (final eating zone), uC THICKCOIL strip thickness at the furnace entrance, mm TMPP2 strip temperature at the heating zone exit, uC TMPP2CNG strip set point temperature at the heating zone exit, uC TMPP1 strip temperature at the heating zone entrance, uC V, Ti, B, N chemical composition of steel, wt-% VELMED strip velocity inside the furnace, m min 21 WIDTHCOIL strip width at the furnace entrance, mm e emissivity Introduction The incorporation of new products into industrial processing plants frequently necessitates changes in the ways in which those processes are controlled or manually adjusted. Adjustment usually takes a long time, can result in errors and defects in products and puts production engineers under additional stress. There is increasing demand in industry for a process model capable of responding correctly to the require- ments not just of product types already processed, but also of new types. The new and classical techniques of data mining (DM) and artificial intelligence (AI) enable past data to be used to develop efficient models capable of improving on previous results. The challenge is to develop overall models that can learn from the past but can still be efficient when faced with new operating conditions in the future. This article compares numerous new and classical DM and AI techniques and their practical application to the development of an overall model capable of precisely predicting furnace heating temperatures on a hot dip galvanising line (HDGL). These comparisons are used 1 EDMANS Group, Departamento de Ingenierı ´a Meca ´ nica, Universidad de La Rioja, Logron ˜ o, Spain 2 GI-TENECO Group, Departamento de Ingenierı ´a Meca ´ nica, Universidad de La Rioja, Logron ˜ o, Spain 3 PMQ Group, Departamento de Ingenierı ´a de Organizacio ´n, ETSII, Universidad Polite ´ cnica de Madrid, Madrid, Spain *Corresponding author, email fjmartin@unirioja.es ß 2010 Institute of Materials, Minerals and Mining Published by Maney on behalf of the Institute Received 21 August 2009; accepted 1 November 2009 DOI 10.1179/030192309X12573371383604 Ironmaking and Steelmaking 2010 VOL 37 NO 3 187